Inverters with Different Loads for Ring Oscillators True Random Number Generator Analysis

Abstract

There have been a multitude of security issues surfacing in hardware security at the moment. The hardware and software systems are exposed to a lot of multi-level attacks in the cryptography aspect. True random number generator (TRNG) can provide an answer to the issues by producing random and unpredictable keys or numbers that can be used in hardware and the transactions of the network. The way these random output are produced are through physical phenomenon that produces entropies that can be sampled by TRNG. This research provides an idea on how the performance of TRNG can be enhanced by creating a more random output. Nanoelectronics such as memristors have been adopted and explored to tackle the issues. The memristor has been implemented in the TRNG designs to analyze the performance results of the ouput. Different loads such as transistors, resistor and memristor have been used in the TRNG to investigate the effect on the performance of the TRNG. The TRNG design uses complementary metal oxide semiconductor (CMOS) technology of $0.18 \boldsymbol{\mu} \mathbf{m}$ and is simulated by LT SPICE IV. The TRNG design in 2nd scenario produces the best results and passed 10 out of 12 of the NIST tests compared to other scenarios.